A primordial and reversible TCA cycle in a facultatively chemolithoautotrophic thermophile

See allHide authors and affiliations

Science  02 Feb 2018:
Vol. 359, Issue 6375, pp. 559-563
DOI: 10.1126/science.aao3407

About-face for citrate synthase

Classically, it is thought that citrate synthase only works in one direction: to catalyze the production of citrate from acetyl coenzyme A and oxaloacetate in the tricarboxylic acid (TCA) cycle. The TCA cycle can run in reverse to cleave citrate and fix carbon dioxide autotrophically, but this was thought to occur only with alternative enzymes, such as citrate lyase. Now Nunoura et al. and Mall et al. have discovered thermophilic bacteria with highly efficient and reversible citrate synthase that requires reduced ferredoxin (see the Perspective by Ragsdale). This function is undetectable by metagenomics, but classical biochemistry filled in the gaps seen between the genome sequences and the phenotypes of the organisms. The direction of catalysis depends on the availability of organic versus inorganic carbon and reflects a flexible bet-hedging strategy for survival in fluctuating environments. In evolutionary terms, this capacity might predate the classical TCA cycle and is likely to occur in a wide range of anaerobic microorganisms.

Science, this issue p. 559, p. 563; see also p. 517


Inorganic carbon fixation is essential to sustain life on Earth, and the reductive tricarboxylic acid (rTCA) cycle is one of the most ancient carbon fixation metabolisms. A combination of genomic, enzymatic, and metabolomic analyses of a deeply branching chemolithotrophic Thermosulfidibacter takaii ABI70S6T revealed a previously unknown reversible TCA cycle whose direction was controlled by the available carbon source(s). Under a chemolithoautotrophic condition, a rTCA cycle occurred with the reverse reaction of citrate synthase (CS) and not with the adenosine 5′-triphosphate–dependent citrate cleavage reactions that had been regarded as essential for the conventional rTCA cycle. Phylometabolic evaluation suggests that the TCA cycle with reversible CS may represent an ancestral mode of the rTCA cycle and raises the possibility of a facultatively chemolithomixotrophic origin of life.

Autotrophic carbon fixation is essential for sustaining life on Earth. The reductive tricarboxylic acid (rTCA) cycle and Wood-Ljungdahl pathway have been recognized as the most ancient carbon fixation pathways to ensure the biosynthesis of the five universal precursors of anabolism; acetyl-CoA (coenzyme A), pyruvate, phosphoenolpyruvate, oxaloacetate, and 2-oxoglutarate (15). However, essential components of the presently recognized rTCA cycle, the adenosine 5′-triphosphate (ATP)–dependent citrate cleavage enzymes ATP citrate lyase (ACL) or citryl-CoA synthetase (CCS)/citryl-CoA lyase (CCL), most likely emerged at a later stage from the domains of citrate synthase (CS) and succinyl-CoA synthetase (fig. S1) (69).

The Aquificae strain Thermosulfidibacter takaii ABI70S6T has been recognized as a chemolithoautotroph (10, 11), but this study revealed its ability for chemolithomixotrophic growth with organic acids including acetate and succinate in the presence of yeast extract (Fig. 1 and supplementary text). Hydrogen oxidation coupled with sulfur reduction was essential for growth. The genome sequence of T. takaii ABI70S6T (~1.8 million base pairs) (supplementary text) was not sufficient to identify the carbon fixation pathway, although the strain was expected to harbor a rTCA cycle because of its close phylogenetic relationship with other deeply branching chemolithoautotrophic Aquificae bacteria (911).

Fig. 1 Growth curves of T. takaii grown under chemolithoautotrophic and chemolithomixotrophic conditions.

A gas mixture of H2 and CO2 (80:20) was used for all cultivation experiments in this study.

On the T. takaii genome, genes related to the (r)TCA cycle are encoded in one gene cluster, and only the isocitrate dehydrogenase (ICDH) gene is found in a different locus (table S1). Surprisingly, however, we could not find genes for ACL and CCS/CCL. Genes with essential motifs or signatures of these enzymes were also absent. Genes for the other known microbial carbon fixation pathways could not be identified. As for potential carbon dioxide fixation enzymes, genes for 2-oxoglutarate:ferredoxin oxidoreductase (OGOR), ICDH, pyruvate:ferredoxin oxidoreductase (POR), pyruvate carboxylase, malate dehydrogenase (malic enzyme), phosphoenolpyruvate carboxykinase (PEPCK), acetyl-CoA carboxylase, and propionyl-CoA carboxylase were present (fig. S2 and table S1).

The activities of these enzymes related to carbon fixation were examined with cell-free extract obtained from T. takaii cells grown chemolithomixotrophically. Among the potential carboxylases, activities of POR, OGOR, ICDH, PEPCK, acetyl-CoA carboxylase, and propionyl-CoA carboxylase were detected, whereas those of malic enzyme and pyruvate carboxylase were not. Most of the enzymatic activities in the rTCA cycle were detected (table S1). ATP-independent, CoA-dependent citrate lyase activity (reverse reaction of CS) was present, but ATP/GTP (guanosine 5′-triphosphate)–citrate lyase and CoA-independent citrate lyase activities were absent (Table 1). Both citrate-generating and citrate-cleaving CS activities were detected in cells grown chemolithoautotrophically. The levels of CS activity in both directions and the malate dehydrogenase (MDH) activity were comparable in T. takaii cells grown chemolithoautotrophically and chemolithomixotrophically, suggesting that the two enzymes are expressed constitutively. Intriguingly, the ATP-independent turnover rate of citrate to oxaloacetate in T. takaii was comparable to the ATP-dependent conversion rate observed in other deeply branching chemolithoautotrophic Aquificae bacteria that utilize ACL or CCS/CCL in the classical rTCA cycle (11) (Table 1). To address the possibility that CS was involved in citrate cleavage in T. takaii, we examined the levels of CS activity in other deeply branching bacteria and compared them to that in T. takaii. As a result, the activity in T. takaii was at least two orders of magnitude higher than those observed in Hydrogenobacter thermophilus and Persephonella marina, which utilize the rTCA cycle for growth. All of the three citrate cleavage activities were not observed in Thermodesulfatator indicus cells, consistent with the absence of genes for CS, ACL and CCS/CCL on its genome. These results suggest that there are no enzymes other than CS, ACL, and CCS/CCL that are involved in citrate cleavage in the deeply branching chemolithoautotrophic bacteria.

Table 1 Enzyme activities of CS, ATP citrate lyase (ACL), citryl-CoA synthetase/citryl CoA lyase (CCS/CCL), and malate dehydrogenase (MDH) in T. takaii cells grown chemolithoautotrophically and/or chemolithomixotrophically along with those of other hydrogenotrophic thermophilic bacteria belonging to deeply branching lineages; Thermovibrio, Hydrogenobacter, and Persephonella in Aquificae, and Thermodesulfatator in Thermodesulfobacteria.

All activities were measured at 70°C. A hyphen (-) indicates not detected; N.T., not tested. The parentheses in ACL and CCS/CCL activity measurements indicate that ATP-dependent activity is negligible compared with ATP-independent CS activity.

View this table:

The CS activity in T. takaii was the highest among the enzyme activities tested. A high CS activity is unlikely to co-occur with carbon fixation pathways using acetyl-CoA as an intermediate such as the archaeal 3-hydroxypropionate/4-hydroxybutyrate (3HP/4HB) and dicarboxylate/4-hydroxybutyrate (DC/HB) cycles, as it would compete in consuming acetyl-CoA (12). In Thermoproteus neutrophilus, which utilizes the DC/HB cycle for CO2 fixation and harbors a TCA cycle, the specific activity of CS was only 2% of that of MDH (13). Enzyme activity measurements in T. takaii also suggested the absence of these archaeal pathways (table S1).

RNA sequencing of T. takaii cells indicated that all of the genes involved in the TCA cycle were transcribed under both chemolithoautotrophic and chemolithomixotrophic (with acetate and/or succinate) growth conditions (table S2). Shotgun proteome analysis of the cell-free extract also confirmed the expression of most of these genes in cells grown chemolithoautotrophically and chemolithomixotrophically with acetate or succinate (table S2).

The carbon fixation pathway of T. takaii was further characterized by isotopologue analysis of protein-derived amino acids in cells grown with 13C-labeled substrates (fig. S3). Genomic information suggests that alanine (Ala), aspartate (Asp), and glutamate (Glu) are synthesized from pyruvate, oxaloacetate, and 2-oxoglutarate by alanine aminotransferase, aspartate aminotransferase, and glutamate dehydrogenase, respectively (fig. S4). The isotopologues of amino acids from cells grown chemolithoautotrophically in the presence of [13C]bicarbonate for 3 hours can be explained by a citrate cleavage reaction and production of [1-13C]acetyl-CoA through the rTCA cycle (fig. S5 and tables S3 and S4). Intriguingly, although anaplerosis is necessary to sustain the level of cycle intermediates and maintain chemolithoautotrophy, the contribution of anaplerotic pathways was not clearly observed. Further studies will be necessary to clarify this issue. Shotgun proteome analysis of the cell-free extract of cells grown chemolithoautotrophically confirmed the presence of pyruvate carboxylase, PEPCK, and malic enzyme (fig. S2), indicating the potential for anaplerosis with these enzymes.

The isotopologues from cells grown chemolithomixotrophically in the presence of succinate were examined with either [2,3-13C]succinate or [13C]bicarbonate. The isotopologue pattern can be explained by a bifurcated TCA cycle (fig. S5 and tables S3 and S4). The isotopologue pattern of Glu indicated the reductive generation of 2-oxoglutarate from succinate via succinyl-CoA. In contrast, that of Ala suggested that pyruvate was formed via decarboxylation of malate or oxaloacetate and not via conversion of acetyl-CoA to pyruvate. Moreover, that of Asp also indicated the dominance of oxidative generation of oxaloacetate from succinate, and a portion was supplied by reduction from succinate via succinyl-CoA, 2-oxoglutarate, isocitrate, and citrate. Therefore, in this case, succinate proceeds in both the oxidative and reductive directions of the cycle, converging at acetyl-CoA via oxaloacetate and pyruvate in the former case and via 2-oxoglutarate, citrate, and citrate cleavage in the latter.

The isotopologue patterns in cells grown with acetate suggested a bifurcated TCA pathway from acetyl-CoA (fig. S5 and and tables S3 and S4). Growth with acetate was examined in the presence of [1-13C]acetate, [1,2-13C]acetate, or [13C]bicarbonate. The isotopologue pattern of Ala suggested that pyruvate was generated from acetyl-CoA (carboxylation), whereas that of Glu suggested the generation of 2-oxoglutarate via oxidation (decarboxylation) of citrate, and not via reduction (carboxylation) of oxaloacetate. The presence of [5-13C] or [4,5-13C2]Glu in the cells grown with [1-13C] or [1,2-13C2]acetate, respectively, implied that a portion of cellular oxaloacetate was supplied from the yeast extract component in the medium. Examination of Asp and Ala indicated that pyruvate was generated from acetyl-CoA derived from acetate and oxaloacetate was formed via carboxylation of PEP or pyruvate. Therefore, in this case, acetate acts as the precursor of acetyl-CoA.

Intriguingly, in the presence of both succinate and acetate with [13C]bicarbonate, signatures of inorganic carbon fixation were not observed in the isotopologues of amino acids (fig. S5 and tables S3 and S4). This implied that the flux with both succinate and acetyl-CoA proceeded in the direction of decarboxylation. This is in contrast to the rTCA cycle with ACL from Chlorobaculum tepidum, which maintains its direction even in the presence of acetate (14). Inorganic carbon fixation examined with [13C]bicarbonate was also absent in cells grown chemolithomixotrophically with yeast extract as the sole organic carbon source. The result suggests that the directions of the TCA cycle in the presence of succinate and/or acetate were not affected by the presence of yeast extract.

A phylogenetic analysis of the CS domain indicated that the CS from T. takaii belonged to a cluster of CSs from deeply branching anaerobic bacteria, Deltaproteobacteria, Bacteroidetes, and eukaryotes (fig. S6). The CSs of Desulfurella acetivorans in Deltaproteobacteria, one of which functioned in the rTCA cycle (15), were also found in this cluster, whereas Deltaproteobacteria is distinct from Aquificae in the small subunit ribosomal RNA gene phylogeny. The phylogenetic topology suggests that both ACL and CCL likely evolved from a lineage of CSs. Among the characterized CSs, the CS from T. takaii was most related to the enzymes from eukaryotes, and there were no marked differences that suggest a bidirectional role for the T. takaii enzyme (fig. S7).

We produced and purified recombinant CS and MDH from T. takaii (Fig. 2). In the case of CS, we measured activity in both the direction of citrate generation and cleavage. With CS alone, only very small amounts of oxaloacetate were formed, and it was difficult to measure the activity using MDH. With an increase in enzyme and longer reaction times, we clearly observed generation of acetyl-CoA by high-performance liquid chromatography (HPLC) analysis (fig. S8A), indicating the occurrence of the citrate-cleaving activity. When we directly coupled the CS activity with the MDH activity, a significant amount of ATP-independent citrate lyase activity was observed, which was dependent on the presence of CS, citrate, and CoA (Table 2 and fig. S9). Production of acetyl-CoA under these conditions was also confirmed by HPLC (fig. S8B). The Michaelis constant (Km) values for citrate (8.6 μM) and CoA (17.4 μM) were extremely low, even when compared to those for previously characterized ATP-citrate lyase enzymes (Table 3). We next examined the citrate-generating activity. Measurements were crried out by quantifying the generation of CoA with varying concentrations of acetyl-CoA or oxaloacetate in the presence of 3000 μM oxaloacetate or 1000 μM acetyl-CoA, respectively (Table 2 and fig. S9). During the analysis at 70°C, we observed thermal degradation of acetyl-CoA to CoA, which was problematic particularly in measurements at low concentrations of oxaloacetate and 1000 μM acetyl-CoA. We thus carried out measurements in the presence of 200 μM acetyl-CoA when varying the concentration of oxaloacetate. The kcat/Km values for the citrate-generating reaction were lower than expected, and this may be due to the thermal degradation of acetyl-CoA in our reactions. However, the high affinity and high kcat/Km value with citrate (and CoA) support our hypothesis that the CS from T. takaii functions in the direction of citrate cleavage in this organism. In the case of MDH, we could carry out kinetic analyses on the reaction from oxaloacetate to malate. The Km value for oxaloacetate was 20.1 ± 6.2 μM. The Vmax value was relatively high (1550 ± 155 μmol min−1 mg−1), and the kcat/Km value was calculated as 43,300 s−1 mM−1 (Table 3 and fig. S10). We cannot judge whether these values are high or not, as there are no MDH proteins that have been characterized from organisms relying on the rTCA cycle for growth. We further examined the overall equilibrium of the CS and MDH reactions converting citrate, CoA, and reduced nicotinamide adenine dinucleotide (NADH) to malate, acetyl-CoA, and NAD+ at 50°C and obtained an equilibrium constant of K = 0.046 ± 0.009 (fig. S11). Analyses of the two enzymes from T. takaii suggest that the enzymes can function in the direction of the rTCA cycle.

Fig. 2 SDS–polyacrylamide gel electrophoresis of recombinant citrate synthase (CS) (TST_0783) and malate dehydrogenase (MDH) (TST_0784) proteins from T. takaii individually expressed in E. coli.

One microgram of protein was loaded onto each lane.

Table 2 Kinetic parameters of citrate synthase (CS) from T. takaii.

View this table:
Table 3 Kinetic parameters of citrate synthase (CS) for citrate and malate dehydrogenase (MDH) for oxaloacetate from T. takaii, and comparison with those of ATP-citrate lyases (ACLs).

View this table:

Accordingly, we conclude that T. takaii constitutively expresses and utilizes a reversible TCA cycle dependent on a bidirectional function of CS as its main carbon fixation pathway. Energetic and kinetic considerations theoretically support this possibility, given a sufficient energy supply via electron bifurcation. The proposed rTCA cycle in T. takaii includes three endergonic reactions: the ATP-independent citrate cleavage reaction (the free energy change under standard conditions δG0 = +37.6 kJ/mol) (16) and the oxoglutarate synthase (OGOR with reduced ferredoxin) (δG0′ = +19 kJ/mol) and pyruvate synthase (POR with reduced ferredoxin) (δG0′ = +19 kJ/mol) reactions (17) (table S5). The overall reductive conversion of succinate to oxaloacetate via 2-oxoglutarate, isocitrate, and citrate is endergonic (δG0 = +55 kJ). The reductive conversion from oxaloacetate to succinate is exergonic; δG0 = –95 kJ when fumarate is reduced by NADH (table S5). The conversion should become even more favorable when reduced ferredoxin is used for fumarate reduction. Several reducing reactions in anaerobic organisms are driven by electron donors with low redox potential, such as reduced ferredoxins. The supply of electron donors with low redox potential, critical in sustaining the driving force of the reaction in the reducing direction, is made possible by flavin-based electron bifurcation (17, 18). Therefore, sufficient depletion of both acetyl-CoA and oxaloacetate and an abundant supply of reducing equivalents via electron bifurcation should allow the cycle to overcome the energetically unfavorable citrate-cleavage reaction of CS. The obligate hydrogenotrophic chemolithotrophy of T. takaii may serve to ensure the sufficient supply of reduced ferredoxin for POR, OGOR, and possibly fumarate reductase. By contrast, when T. takaii is grown chemolithomixotrophically with acetate, isotopologue analysis indicated that 2-oxoglutarate is generated solely by oxidative conversion from acetyl-CoA and oxaloacetate, driven by the exergonic CS reaction (fig. S5). When T. takaii is grown chemolithomixotrophically with succinate, isotopologue analysis indicated that a major portion of the oxaloacetate is generated by oxidative conversion of succinate. A proton motive force maintained by hydrogen oxidation coupled with S0 reduction should be necessary for the endergonic oxidation of succinate in these cells (fig. S5), as in the case of Desulfuromonas acetioxidans (19).

The rTCA cycle is believed to be one of the most ancient carbon fixation pathways on Earth. Our results suggest that the ancient rTCA cycle in the early life on Earth did not necessarily require an ACL or CCS/CCL and originated from a bidirectional CS-dependent TCA cycle as identified in T. takaii. Identification of this previously unknown reversible TCA cycle and its functional regulation by the availability of carbon sources provides new insights that will help us understand the evolution of carbon fixation and central metabolisms in the earliest organisms on Earth. There is a long history of debate over heterotrophic versus autotrophic origin-of-life scenarios (20). Recently, abiotic synthesis of rTCA cycle components was observed in the presence of sulfate radicals (21). The directions of the individual conversions depended on the reaction, as there was no external energy input. When energy supplies are sufficient, the metabolic commitment toward autotrophy and heterotrophy driven by the primordial TCA cycle in T. takaii is flexible and would be controlled by environmental conditions such as availability of carbon sources. It has been pointed out that the core anabolism is universal at the ecosystem level, and the configuration of autotrophy or heterotrophy is an ecological distinction (22). The reversible TCA cycle, as seen in T. takaii, would meet these anabolic requirements and serve both autotrophic and heterotrophic lifestyles. The versatile TCA cycle would allow the earliest possible chemolithotrophic forms of life, with their limited metabolic network and catalytic enzyme (gene) resources, to make the most of what is available in their environment, such as in deep-sea hydrothermal vents. Thus, ancestral forms of life may have originated in the form of facultative autotrophs or mixotrophs. Further investigations into the possibilities of a facultative chemolithomixotrophic origin of life will provide key insights into proposing more likely scenarios for the origin of life on Earth.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S11

Tables S1 to S5

References (2851)

References and notes

Acknowledgments: All methods and additional figures and text are available in the supplementary materials. T.N. designed the study; T.N. and K.M. contributed in cultivation analyses; T.N., Y.C., K.Y., M.M., and N.O. conducted metabolomics; T.H., Y.K., A.S., N.F., and Y.T. analyzed genome and/or transcriptome; S.S. conducted proteomics; T.N., T.Sa., and H.A. conducted enzymatic assays; and R.I, T.Su., T.Sa., and H.A contributed in expression and characterization of recombinant enzymes. T.N., Y.C., T.Sa., H.A., and K.T. wrote the manuscript. All data and code required to understand and assess the conclusions of this research are available in the main text, supplementary materials, and via the following repositories. The genome sequence and transcriptome data set have been deposited in the DNA Data Bank of Japan, European Molecular Biology Laboratory, and GenBank with the accession numbers AP013035 and DRA004184, respectively. The proteome data set is available in jPOST with the accession numbers JPST000334 and PXD007930 (ProteomeXchange). The phylogenetic tree and alignment of CS domain are available in TreeBASE with the study ID S21638.
View Abstract

Navigate This Article